Zebra finch (Taeniopygia guttata) shift toward aerodynamically efficient flight kinematics in response to an artificial load

Abstract

We investigated the effect of an added mass emulating a transmitter on the flight kinematics of zebra finches (Taeniopygia guttata), both to identify proximal effects of loading and to test fundamental questions regarding the intermittent flight of this species. Zebra finch, along with many species of relatively small birds, exhibit flap-bounding, wherein the bird alternates periods of flapping with flexed-wing bounds. Mathematical modeling suggests that flap-bounding is less aerodynamically efficient than continuous flapping, except in limited circumstances. This has prompted the introduction of two major hypotheses for flap-bounding - the 'fixed-gear' and 'cost of muscle activation/deactivation' hypotheses - based on intrinsic properties of muscle. We equipped zebra finches flying at 10 m s^-1 with a transmitter-like load to determine if their response was consistent with the predictions of these hypotheses. Loading caused finches to diverge significantly from their unloaded wingbeat kinematics. Researchers should carefully consider whether these effects impact traits of interest when planning telemetry studies to ensure that tagged individuals can reasonably be considered representative of the overall population. In response to loading, average wingbeat amplitude and angular velocity decreased, inconsistent with the predictions of the fixed-gear hypothesis. If we assume that finches maintained muscular efficiency, the reduction in amplitude is inconsistent with the cost of the muscle activation/deactivation hypothesis. However, we interpret the reduction in wingbeat amplitude and increase in the proportion of time spent flapping as evidence that loaded finches opted to increase their aerodynamic efficiency - a response which is consistent with the latter hypothesis.

Keywords: Flap-bounding; Intermittent flight; Kinematics; Zebra finch.

Fig. 1.

Kinematic parameters describing the wingbeats of zebra finch (T. guttata) when equipped with a load weighing 10% of their body mass (‘Loaded’) versus the kinematics of wingbeats when those same birds were unweighted (‘Unloaded’). Each panel displays the distribution of parameter values measured across all individuals (n=5) with the unweighted average for each condition indicated by a vertical line of the same color. Linear mixed-effects models for each parameter were tested for a significant effect of loading using the KRmodcomp function in the pbkrtest package in R. (A) Loaded wingbeats were of significantly reduced amplitude (*P<0.01, n=144) relative to loaded wingbeats, from an unloaded average of 109.5±19.5° to 101.5±20.7° under load. (B) Loaded wingbeats were of significantly reduced downstroke velocity (**P<0.001, n=142), from an unloaded average of 123.1±21.0 rad s⁻¹ to 109.9±21.6 rad s⁻¹ under load. (C) Loaded wingbeats were of significantly increased angle of incidence (*P<0.01, n=126), from an unloaded average of 11.0±6.4° to 14.0±6.2° under load. (D) There was no significant effect of loading on distal angle of incidence (P=0.056, n=118).

Fig. 2.

Birds equipped with a payload weighing 10% of their body mass exhibited wingbeats (n=128) of significantly reduced total aerodynamic power, P_aero, relative to their unloaded wingbeats (P=0.031; linear mixed-effects model, Kenward-Rogers-adjusted P-value). The color/marker of each point indicates the individual finch from which that wingbeat was sampled, while the violins represent the distribution of the pooled wingbeats in each treatment. Average P_aero was reduced from 3.03±0.82 W when unloaded to 2.74±0.69 W under load. The scatter of the points within each individual highlight that the wingbeats of zebra finch cannot be considered ‘fixed’, contrary to the prediction of the fixed-gear hypothesis.